The Bayesian information criterion (BIC) has become a popular criterion for model selection in recent years. The BIC is intended to provide a measure of the weight of evidence favoring one model over another, or Bayes factor. It has, however, some important drawbacks that are not widely recognized. First, Bayes factors depend on prior beliefs about the expected distribution of parameter values, and there is no guarantee that the Bayes factor implied by the BIC will be close to one calculated from a prior distribution that an observer would actually regard as appropriate. Second, to obtain the Bayes factors that follow from the BIC, investigators would have to vary their prior distributions depending on the marginal distributions of the variables and the nature of the hypothesis. Such variations seem unwarranted in principle and tend to make the BIC inclined to favor excessively simple models in practice. These points are illustrated by the analysis of several examples, and alternatives to use of the BIC are discussed.
Whereas many infectious diseases are spread through casual contact and contagion,HIV transmission results from risk behaviors that involve close and often intimate contact. As a result, the transmission of HIV is structured by the social relationships within which these contacts are embedded. Hence, social network analysis is especially suitable for understanding the AIDS epidemic. This paper reports the results of a field experiment that compares a network-based HIV prevention intervention, termed a "Peer-Driven Intervention" (PDl), with the standard form of streetbased outreach intervention.The results suggest that the network intervention outperforms the standard approach with respect to number of people accessed, reductions in self-reported levels of HIV risk behavior and cost. Finally, the analysis focuses on the network structure of drug injectors and discusses the implications of these structures for understanding both the spread of HIV through social networks and the design of HIV-prevention interventions.The results show that certain network features, including geographically extensive networks and an abundance of ties across ethnic boundaries, genders, ages and drug preferences, can further the spread of HIV.Ironically, these are also the network features that increase the effectiveness of network-based HIV-prevention interventions.Thus, we show that network interventions work best precisely when they are most needed, that is, when network structures facilitate the spread of HIV.
This study examines the relationship between educational attainment and various dimensions of religiosity. On the basis of a network closure argument, we hypothesize that the relationship between education and religiosity varies by religious tradition. Analyzing data from the 1972-2006 General Social Survey, we found that educational attainment predicted increased attendance at religious services, decreased levels of prayer, increased inclination to view the Bible as a book of fables, and decreased inclination to view the Bible as the literal word of God. These relationships, however, significantly interacted with religious tradition. Increased education largely resulted in greater religiosity among evangelical Protestants, black Protestants, and Catholics but not among mainline Protestants and the nonaffiliated. Overall, this study shows that education does not uniformly decrease religiosity and highlights the importance of considering religious tradition in future research.
This study considers the relationship between morale and workplace productivity in a representative sample of Australian workplaces. It focuses on three questions: the shape of the relationship, whether the effects of morale are contingent on other factors, and the paths by which any effect takes place. The results show that morale influences productivity in an approximately linear fashion. The effect of morale on productivity appears to be larger when management regards product quality as important and attempts to develop a corporate ethic and culture. Morale is associated with greater work effort, but the relationship between work effort and productivity becomes stronger at higher levels of morale. Thus, part of the influence of morale on productivity is a matter of increasing the effectiveness of workers’ efforts.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.